Forests have an essential carbon function, removing each year about 20% of global anthropogenic CO2 emissions. Models have been designed to quantify the distribution of CO2 fluxes and ecosystem carbon dynamics at the global scale, but these global vegetation models usually do not account for management and stand structure. In most cases - and in particular in the biosphere component of the IPSL ESM model (ORCHIDEE) - forests are represented as an average single tree that has reached a near-equilibrium state. This coarse representation has so far impaired the capacity of global models to correctly simulate the interactions between managed forests and the climate system. This presentation describes a new module that explicitly simulates forest management, stand structure, and tree mortality within ORCHIDEE. Forest representation is detailed from an "average tree" to an "average stand", with an explicit distribution in diameter classes that varies with stand age, location and climate. A set of rules are defined to calculate the critical tree density, intensity and frequencies of thinning, etc... Through a sensitivity analysis, the reaction of the model to changes in management and physiological parameters is presented. The model¿s fit to forest inventory data and yield tables is assessed. Eventually, the explicit simulation of forest management is shown to have an important impact on the simulation of biochemical cycles, both in terms of carbon fluxes and stocks.